A line based pose representation for human action recognition

In this paper, we utilize a line based pose representation to recognize human actions in videos. We represent the pose in each frame by employing a collection of line-pairs, so that limb and joint movements are better described and the geometrical relationships among the lines forming the human figure are captured. We contribute to the literature by proposing a new method that matches line-pairs of two poses to compute the similarity between them. Moreover, to encapsulate the global motion information of a pose sequence, we introduce line-flow histograms, which are extracted by matching line segments in consecutive frames. Experimental results on Weizmann and KTH datasets emphasize the power of our pose representation, and show the effectiveness of using pose ordering and line-flow histograms together in grasping the nature of an action and distinguishing one from the others.     

A Unified Approach to Iconic Indexing, Retrieval, and Maintenance of Spatial Relationships in Image Databases

In this paper we propose a unified iconic indexing, the generalized combined 2D string representation, for images in image databases. Each 2D image is modelled as a generalized extended pseudo-symbolic picture, which has the GEP-2D string representation. We present an efficient algorithm to generate the GEP-2D string representation for each 2D image. We also show how to maintain the complete information about the absolute spatial relationships in the image. Our proposed iconic indexing combines both the GEP-2D string representation and the usual 2D string representation to capture absolute and relative spatial relationships in the image. The result is better representation of spatial relationships in image databases. These results extend our earlier work on a large class of 2D scenes, the extended pseudo-symbolic pictures. Picture retrieval by generalized combined 2D strings is discussed. Our approach can also be easily formulated for 3D images.     

Statistical bias in 3-D reconstruction from a monocular video.

The present state-of-the-art in computing the error statistics in three-dimensional (3-D) reconstruction from video concentrates on estimating the error covariance. A different source of error which has not received much attention is the fact that the reconstruction estimates are often significantly statistically biased. In this paper, we derive a precise expression for the bias in the depth estimate, based on the continuous (differentiable) version of structure from motion (SfM). Many SfM algorithms, or certain portions of them, can be posed in a linear least-squares (LS) framework Ax = b. Examples include initialization procedures for bundle adjustment or algorithms that alternately estimate depth and camera motion. It is a well-known fact that the LS estimate is biased if the system matrix A is noisy. In SfM, the matrix A contains point correspondences, which are always difficult to obtain precisely; thus, it is expected that the structure and motion estimates in such a formulation of the problem would be biased. Existing results on the minimum achievable variance of the SfM estimator are extended by deriving a generalized Cramer-Rao lower bound. A detailed analysis of the effect of various camera motion parameters on the bias is presented. We conclude by presenting the effect of bias compensation on reconstructing 3-D face models from rendered images.     

多导体传输线的邻近效应和分布参数研究

随着多导体传输线内各导体之间间距的减小, 导体之间的近邻效应对传输线的分布参数和传输特性的影响越来越大.为此, 我们针对三种典型的传输线结构, 分别建立了基于矢势有限元方法分析的多导体传输线的模型, 并分析了近邻效应对磁通密度和分布电感的影响.利用提出的方法计算了同轴传输线的单位长度分布电感, 并将它与采用解析方法得到的结果进行比较来证明该方法的正确性.计算双线传输线在不同间距时的单位长度电感, 与理论分析得到的结果相比较验证了导线间距越小, 近邻效应对单位长度电感的影响越大.最后, 计算考虑了近邻效应的耦合微带线的电感矩阵, 并将它与其他不考虑近邻效应的方法得到的结果相比较, 说明近邻效应对传输线电感矩阵的影响.     

Spatio-temporal segmentation of image sequences for object-oriented low bit-rate image coding

This paper describes a new method of segmentation of time-varying image sequences whose goal is object-oriented image coding. The segmentation represents a partition of each frame of the sequence into a set of regions which are homogeneous with regard to motion criterion. The region borders correspond to spatial contours of objects in the frame. Each spatio-temporal region is characterized by its temporal component, which is a model-dependent vector of motion parameters, and a structural component representing the polygonal approximation of the spatial contour of the region.

The construction of spatio-temporal segmentation includes two phases: the initialization step and temporal tracking. The initialization step is based on the spatial segmentation of the first frame of the sequence. Then homogeneous spatial regions are merged through motion estimation in accordance with a motion-based criterion. The temporal tracking consists of the projection of the segmentation along the time axis, and its adjustment. Special attention is paid to the processing of occlusions.

A predictive coding scheme is proposed which is based on the temporal coherence of the segmentation. This scheme is promising for a low bit-rate image compression.

The results for teleconference and TV sequences show the high visual quality of reconstructed only by prediction images. Moreover, the bit-rates for motion coding are very low: from 0.002 to 0.007 bit/pixel for teleconference sequence and from 0.004 to 0.021 bit/pixel for complex TV sequence. A scheme for encoding of the structural information is proposed which requires 0.083 – 0.17 bit per pixel depending on the content of the sequence.     

Signal representation in the combined phase-spatial space:reconstruction and criteria for uniqueness

Motivated by the major roles of the spectral phase in signal structure and by previous results on the importance of localized representation in images, we develop conditions for unique representation of a signal by a combination of its spectral (Fourier) phase and spatial samples. Methods for signal reconstruction from various combinations of spatial and phase information are introduced. It is concluded that most of the previous results in the area of phase-only representation and reconstruction are special cases of the general approach presented here     

A New Approach to Motion Pattern Recognition and Its Application to Optical Flow Estimation

In this paper, a new methodology for extracting motion patterns is applied to optical flow estimation in the presence of multiple motions. The proposed approximation deals with the problem in two stages. In the first one, the most important motions are segmented; in the second one, the optical flow is estimated on the basis of the motions detected in the previous stage. To extract relevant motions, a new approach based on a spatio-temporal filtering is presented. The approach groups together parts of a moving object that have been separated into various filter responses because of the object's spatial structure, thereby avoiding the spatial dependency problem associated with a representation based on spatio-temporal filters. The proposed model, therefore, generates one "motion pattern" for each motion detected in the sequence. To obtain an optical flow estimation, which is able to represent multiple velocities, the gradient constraint is applied to the output of each filter so that multiple estimations of the velocity at the same location may be obtained. For each "motion pattern" detected in the previous stage, the velocities at a given point corresponding to the same motion are then combined using a probabilistic approach. In the application to optical flow estimation, the use of "motion patterns" allows multiple velocities to be represented, while the combination of estimations from different filters helps reduce the aperture problem. This technique is illustrated on real and simulated data sets, including sequences with occlusion and transparencies     

Fixed and Adaptive Predictors for Hybrid Predictive/Transform Coding

Hybrid predictive/transform coding is studied. The usual formulation is to first apply a unitary transform and then code the transform coefficients with independent DPCM coders, i.e., the prediction is performed in the transform domain. This structure is compared to spatial domain prediction, where a difference signal is formed in the spatial domain and then coded by a transform coder. A linear spatial domain predictor which minimizes the mean square prediction error also minimizes the mean square of each transform coefficient. The two structures are equivalent if the transform domain prediction scheme is extended to a more general predictor. Hence, the structure that gives the easiest implementation can be chosen. The spatial domain structure is preferred for motion compensation and for line interlaced video signals. Interframe hybrid coding experiments are performed on interlaced videophone scenes using an adaptive transform coder. Motion compensation gives a rate reduction of 25-35 percent compared to frame difference prediction with the same mean square error. The subjective advantage is even greater, since the "dirty window" effect is not present with motion compensation. It is important to perform the motion estimation with fractional pel accuracy. Field coding with a switched predictor using previous field in moving areas is an interesting alternative to frame coding with frame difference prediction.     

Study of the DC biasing effect on insertion losses in high-frequency interconnections

The paper deals with an experimental investigation of the behavior of high-frequency Si/SiO₂/Al based interconnects when an extra DC bias voltage is applied, by means of which the conductor line changes the surface properties of the semiconductor substrate. By superposing a DC bias to the high-speed signal applied to the line, the insertion losses caused by the semiconductor substrate show a significant decrease over the observed frequency range. In order to study this effect a number of test samples containing several microstrip asymmetric transmission lines were prepared and measured. The obtained results suggest a way of controlling the performance and energy propagation of interconnects on semiconductor substrates. The observed effect can be successfully applied in high-speed blocks with tunable parameters.     

基于轴向空间注意力和中间融合表示的单图像三维重建模型

提出一种基于注意力和中间融合表示的三维重建模型，旨在重建具有精细化结构的三维模型。该方法利用轴向空间注意力机制学习不同方向的信息，将其嵌入编码器中以捕获局部结构特征；并基于双流网络推测深度图和三维平均形状以设计中间融合表示模块，该模块能够有效地融合可见表面细节信息，从而更好地描绘对象的三维空间结构。实验结果表明：所提出的轴向空间注意力机制和中间融合表示模块增强了特征提取的能力，IoU和F-score比PixVox++分别提升了1.3%和0.4%,三维重建效果更优。     

基于稀疏表示和道路辅助的单幅SAR图像运动目标检测方法

本文提出一种利用单幅SAR(Synthetic Aperture Radar)图像实现运动目标检测的方法.首先提出一种基于压缩感知的SAR图像道路检测算法:根据SAR图像中道路的特点,使用模糊C均值方法将图像进行模糊分类,获得大致的道路区域,然后利用Hough变换域的稀疏性,用压缩感知精确定位图像中的道路信息.其次利用图像稀疏表示的方法对运动目标进行检测:不同速度运动目标的散焦量和距离单元跨越不同,由此生成样本图像,继而构造超完备字典.将待测图像分块,并计算子图像在字典下的稀疏系数,检测并匹配出运动目标的速度参数.最后,结合已检测出的道路辅助信息,消除多普勒模糊影响,剔除虚假的运动目标,并对运动目标速度参数进行校正.实验结果证明了所提方法的有效性.     

一种掌纹纹线结构特征的描述和匹配方法

掌纹的纹线结构特征明显,稳定且具有足够的辨别信息。由于掌纹线很不规则,很难用数学表达式的方式对其结构进行精确描述。该文提出一种新的用于掌纹识别的纹线结构特征描述和匹配方法。该方法应用改进的注水算法从二值的纹线图像上提取纹线的统计结构特征;标记纹线方向,提取纹线的方向结构特征;将掌纹线的全局统计结构特征和局部细节方向结构特征整合为一个特征向量描述掌纹纹线的结构特征;最后定义一个加权距离衡量不同纹线结构特征之间的相似度。实验表明,该方法不仅具有较好的匹配性能,而且可以获得较高的识别精度。     

Line detection in images through regularized Hough transform.

The problem of determining the location and orientation of straight lines in images is of great importance in the fields of computer vision and image processing. Traditionally the Hough transform, (a special case of the Radon transform) has been widely used to solve this problem for binary images. In this paper, we pose the problem of detecting straight lines in gray-scale images as an inverse problem. Our formulation is based on use of the inverse Radon operator, which relates the parameters determining the location and orientation of the lines in the image to the noisy input image. The advantage of this formulation is that we can then approach the problem of line detection within a regularization framework and enhance the performance of the Hough-based line detector through the incorporation of prior information in the form of regularization. We discuss the type of regularizers that are useful for this problem and derive efficient computational schemes to solve the resulting optimization problems enabling their use in large applications. Finally, we show how our new approach can be alternatively viewed as one of finding an optimal representation of the noisy image in terms of elements chosen from a dictionary of lines. This interpretation relates the problem of Hough-based line finding to the body of work on adaptive signal representation.     

Synergizing spatial and temporal texture

Temporal texture accounts for a large proportion of motion commonly experienced in the visual world. Current temporal texture techniques extract primarily motion-based features for recognition. We propose a representation where both the spatial and the temporal aspects of texture are coupled together. Such a representation has the advantages of improving efficiency as well as retaining both spatial and temporal semantics. Flow measurements form the basis of our representation. The magnitudes and directions of the normal flow are mapped as spatiotemporal textures. These textures are then aggregated over time and are subsequently analyzed by classical texture analysis tools. Such aggregation traces the history of a motion which can be useful in the understanding of motion types. By providing a spatiotemporal analysis, our approach gains several advantages over previous implementations. The strength of our approach was demonstrated in a series of experiments, including classification and comparisons with other algorithms.     

基于随机Hough变换的三维集中式航迹起始算法

本文提出了一种用于三维空间直线航迹起始的集中式Hough变换法.基于三维直线的Roberts表示法,设计了Roberts参数相似性判决的假设检验方法.通过对多传感器点迹的随机采样,计算每次采样确定的直线Roberts参数,然后以相似性测度实现Roberts参数的合并,当参数累积超过规定的门限后能可靠地起始一条直线航迹.为了减少无意义的随机采样和不合理的Roberts参数合并,在算法实时过程中适时地检查目标运动速度限制,加快了采样速度,减少了航迹误起始的概率.通过仿真,该算法能适应多传感器多目标环境中传感器数据更新率不同和密集杂波、强干扰的复杂情况.     

Image segmentation and selective smoothing by using Mumford-Shah model.

Recently, Chan and Vese developed an active contour model for image segmentation and smoothing by using piecewise constant and smooth representation of an image. Tsai et al. also independently developed a segmentation and smoothing method similar to the Chan and Vese piecewise smooth approach. These models are active contours based on the Mumford-Shah variational approach and the level-set method. In this paper, we develop a new hierarchical method which has many advantages compared to the Chan and Vese multiphase active contour models. First, unlike previous works, the curve evolution partial differential equations (PDEs) for different level-set functions are decoupled. Each curve evolution PDE is the equation of motion of just one level-set function, and different level-set equations of motion are solved in a hierarchy. This decoupling of the motion equations of the level-set functions speeds up the segmentation process significantly. Second, because of the coupling of the curve evolution equations associated with different level-set functions, the initialization of the level sets in Chan and Vese's method is difficult to handle. In fact, different initial conditions may produce completely different results. The hierarchical method proposed in this paper can avoid the problem due to the choice of initial conditions. Third, in this paper, we use the diffusion equation for denoising. This method, therefore, can deal with very noisy images. In general, our method is fast, flexible, not sensitive to the choice of initial conditions, and produces very good results.     

Adaptive Morphological Representation of Signals: Polynomial and Wavelet Methods

In this paper, we propose a novel signal representation based on mathematical morphology, and with it develop representations analogous to the polynomial transform and the bank-of-filters implementation of the wavelet representation. The geometric decomposition of a signal is achieved by separating it into analysis frames and applying mathematical morphological operators with adaptive structuring functions to each frame. The adaptation parameters are found by solving iteratively nonlinear equations that result from constraining the morphological results to achieve optimal fitting. If the structuring functions are derived from real-valued orthogonal polynomials defined on a window, the representation is analogous to the polynomial transform. Using a morphological interpolation, we derive a pyramid-like structure to decompose a signal into gross and fine information components, at different scales, just as in the wavelet transformation. Non-linear morphological operators reduce the computational complexity of the proposed representations. Although these representations are easily extended to two-dimensions, one needs to consider the non–unique ordering of the structuring functions, and the different sampling, decimation and interpolation procedures in two-dimensions. The application of our procedures is mainly in image data compression, but they could also used in object identification. We illustrate our representations by means of one- and two-dimensional examples.     

Multiple motion segmentation with level sets

Segmentation of motion in an image sequence is one of the most challenging problems in image processing, while at the same time one that finds numerous applications. To date, a wealth of approaches to motion segmentation have been proposed. Many of them suffer from the local nature of the models used. Global models, such as those based on Markov random fields, perform, in general, better. In this paper, we propose a new approach to motion segmentation that is based on a global model. The novelty of the approach is twofold. First, inspired by recent work of other researchers we formulate the problem as that of region competition, but we solve it using the level set methodology. The key features of a level set representation, as compared to active contours, often used in this context, are its ability to handle variations in the topology of the segmentation and its numerical stability. The second novelty of the paper is the formulation in which, unlike in many other motion segmentation algorithms, we do not use intensity boundaries as an accessory; the segmentation is purely based on motion. This permits accurate estimation of motion boundaries of an object even when its intensity boundaries are hardly visible. Since occasionally intensity boundaries may prove beneficial, we extend the formulation to account for the coincidence of motion and intensity boundaries. In addition, we generalize the approach to multiple motions. We discuss possible discretizations of the evolution (PDE) equations and we give details of an initialization scheme so that the results could be duplicated. We show numerous experimental results for various formulations on natural images with either synthetic or natural motion.